Hi,
I have translated the authconfig/po/gu.po file using the tool called Kbabel (which generally takes care of all kind of syntax errors) and tried to submit it through transifex. But, I got the following error message: "Error: The PO file you submitted doesn't pass the checks for correctness. Please run msgfmt -c on it and correct the errors before submitting.". So, I did "msgfmt -c" on the file I tried to submit, which is perfect as per the suggested command.
msgfmt -c
Can someone please fix this issue?
Let me know, if you require any further information about the issue?
Thanks! Ankit
On Wed, 27 Aug 2008, Dimitris Glezos wrote:
On Wed, Aug 27, 2008 at 4:11 PM, Mike McGrath mmcgrath@redhat.com wrote: On Wed, 27 Aug 2008, Dimitris Glezos wrote: On Wed, Aug 27, 2008 at 3:10 PM, Asgeir Frimannsson asgeirf@gmail.com wrote: Hey guys, it seems fedora's transifex is not working. Always getting msgfmt check errors when submitting. Any clue what could be wrong? (CCing Ricky and Mike) Some of the servers were rebuilt, and maybe the Tx source has changed. We had a few files modified locally on app4 which weren't pushed in the RPM, maybe we need to get them back from a backup. We don't back up source and binary files on the filesystem. They're in RPM. We only backup data. Were the changes on app4 upstream in the git repo? Is there a newer version we should be using? The changes weren't pushed in the upstream hg repo or in the RPM. I couldn't reproduce some of the problems at the time (FAS2, move to RHEL server). I think ricky did attach some of the fixes on tickets: https://fedorahosted.org/transifex/query?status=accepted&status=assigned&status=new&status=reopened&reporter=~ricky&order=priority&col=id&col=summary&col=status&col=owner&col=type&col=priority&col=milestone There's a new version available, 0.3.1, which we could push: http://groups.google.com/group/transifex-devel/browse_thread/thread/f39d1ea0f967bab3 I think the DB might not require migration, or if it does, it will be a trivial one: http://transifex.org/changeset?old_path=%2F&old=222%3Aaf85d5667720&new_path=%2F&new=348%3Af5b058c065e1#file53 Mike, how difficult do you think it would be to tweak the Tx's spec to be packaged in Fedora? Wouldn't koji/bodhi make it easier to build and push updates or would it require weird local rpmbuild environments too?
On Wed, Aug 27, 2008 at 4:11 PM, Mike McGrath mmcgrath@redhat.com wrote:
On Wed, 27 Aug 2008, Dimitris Glezos wrote: On Wed, Aug 27, 2008 at 3:10 PM, Asgeir Frimannsson asgeirf@gmail.com wrote: Hey guys, it seems fedora's transifex is not working. Always getting msgfmt check errors when submitting. Any clue what could be wrong? (CCing Ricky and Mike) Some of the servers were rebuilt, and maybe the Tx source has changed. We had a few files modified locally on app4 which weren't pushed in the RPM, maybe we need to get them back from a backup. We don't back up source and binary files on the filesystem. They're in RPM. We only backup data. Were the changes on app4 upstream in the git repo? Is there a newer version we should be using?
On Wed, Aug 27, 2008 at 3:10 PM, Asgeir Frimannsson asgeirf@gmail.com wrote: Hey guys, it seems fedora's transifex is not working. Always getting msgfmt check errors when submitting. Any clue what could be wrong? (CCing Ricky and Mike) Some of the servers were rebuilt, and maybe the Tx source has changed. We had a few files modified locally on app4 which weren't pushed in the RPM, maybe we need to get them back from a backup.
On Wed, Aug 27, 2008 at 3:10 PM, Asgeir Frimannsson asgeirf@gmail.com wrote:
Hey guys, it seems fedora's transifex is not working. Always getting msgfmt check errors when submitting. Any clue what could be wrong?
Hey guys,
it seems fedora's transifex is not working. Always getting msgfmt check errors when submitting. Any clue what could be wrong?
(CCing Ricky and Mike)
Some of the servers were rebuilt, and maybe the Tx source has changed. We had a few files modified locally on app4 which weren't pushed in the RPM, maybe we need to get them back from a backup.
We don't back up source and binary files on the filesystem. They're in RPM. We only backup data. Were the changes on app4 upstream in the git repo? Is there a newer version we should be using?
The changes weren't pushed in the upstream hg repo or in the RPM. I couldn't reproduce some of the problems at the time (FAS2, move to RHEL server). I think ricky did attach some of the fixes on tickets:
https://fedorahosted.org/transifex/query?status=accepted&status=assigned&status=new&status=reopened&reporter=~ricky&order=priority&col=id&col=summary&col=status&col=owner&col=type&col=priority&col=milestone
There's a new version available, 0.3.1, which we could push:
http://groups.google.com/group/transifex-devel/browse_thread/thread/f39d1ea0f967bab3
I think the DB might not require migration, or if it does, it will be a trivial one:
http://transifex.org/changeset?old_path=%2F&old=222%3Aaf85d5667720&new_path=%2F&new=348%3Af5b058c065e1#file53
Mike, how difficult do you think it would be to tweak the Tx's spec to be packaged in Fedora? Wouldn't koji/bodhi make it easier to build and push updates or would it require weird local rpmbuild environments too?
Shouldn't be too difficult. It'd be best to just get transifex in Fedora so we can just yum update it when a new package comes out.
So there's a couple of issues here.
1) tracking upstream for updates 2) fixing transifex right now
the long term for 1) is something we certainly need to do but will that happen quickly enough for 2) ?
-Mike
I just temporarily disabled msgfmt checks, so submissions should work until we fix this.
Everyone should now CHECK MANUALLY every file about to be submitted using msgfmt -c myfile.po.
msgfmt -c myfile.po
Yes, Transifex is submitting fine now.
Should we close this ticket and leave the msgfmt checks issue until we find a good way to push transifex updates live?
Working for me too!
Thx guys!
I'd say to leave this open until we fix the bug itself (msgfmt errors). We should really enable the checks as soon as possible, otherwise developers will start throwing stones to us.
Did anyone get a chance to take a look at this? We should enable this asap cause some PO files break the code and developers (eg. anaconda) complain about it.
I don't have high knowlege here but I'll see if I can figure out if it's a version mismatch in msgformat or unexpected return code or something easy.
Fixed (thanks toshio) and tested.
Patch will be sent upstream.
Login to comment on this ticket.